Why is proline hydroxylated？

[Anonymous]

Dear Sir/Madam,

I used the relax.linuxgccrelease of Rosetta to Relax my homology model generated by comparative modeling of Rosetta.

The command was
relax.linuxgccrelease @relax.options >& relax.log
The options were:
database /home/g/guanglin/pfs/software/rosetta_akka_S/rosetta_database
-relax:fast
-in:file:s SCESA_com.pdb
-in:file:fullatom
-out:nstruct 2000
-out:file:silent SCESA_com_relax.out
-out:file:silent_struct_type binary
-out:file:fullatom
-out:file:scorefile relax.score

The relaxing work went smoothly. However, when I extracted the pdbs and examined them with PyMol, I found that the prolines were hydroxylated and their residues names were changed to be OAZ or HYP. When I checked the log file, it had the following information:
core.io.pdb.file_data: [ WARNING ] discarding 7 atoms at position 1 in file SCESA_com.pdb. Best match rsd_type: PRO_p:NtermProteinFull_p:pro_hydroxylated_case1
…
core.io.pdb.file_data: [ WARNING ] discarding 5 atoms at position 13 in file SCESA_com.pdb. Best match rsd_type: PRO_p:pro_hydroxylated_case1
..
core.io.pdb.file_data: [ WARNING ] discarding 5 atoms at position 30 in file SCESA_com.pdb. Best match rsd_type: PRO_p:pro_hydroxylated_case1
core.io.pdb.file_data: [ WARNING ] discarding 5 atoms at position 31 in file SCESA_com.pdb. Best match rsd_type: PRO_p:pro_hydroxylated_case1
The residues types were changed to OAZ or HYP.

This is a very strange strange problem. Why would prolines be hydroxylated? Can someone help me with this?

Thank you!

[Anonymous]

Rosetta uses a heuristic method based on the presence/absence of atoms to figure out if it should be viewing a residue as modified. Sometimes that heuristic can go wrong.

My guess is that the reason it’s not liking your prolines as prolines is due to the discarded atoms on the residues. Take a look at the heavy atom names of your proline residues in the input PDB. Are they conforming to PDB standards? More particularly, do they match the atom naming conventions used in rosetta_database/chemical/residue_type_sets/fa_standard/residue_types/l-caa/PRO.params?

Another thing to check is the occupancy column. Depending on commandline settings, Rosetta can discard atoms with zero values in the occupancy column. (The commandline flag “-ignore_zero_occupancy false” can reverse this behavior.)

[Anonymous]

Thanks, I think it might be the heuristic reason… The atom names and occupancies are normal.
For the first proline, in the input PDB:
ATOM 1 N PRO A 1 -3.139 -15.719 0.289 1.00 0.00 N1+
ATOM 2 CA PRO A 1 -4.096 -14.883 -0.427 1.00 0.00 C
ATOM 3 C PRO A 1 -3.392 -13.770 -1.191 1.00 0.00 C
ATOM 4 O PRO A 1 -2.229 -13.463 -0.929 1.00 0.00 O
ATOM 5 CB PRO A 1 -5.012 -14.338 0.674 1.00 0.00 C
ATOM 6 CG PRO A 1 -4.139 -14.261 1.879 1.00 0.00 C
ATOM 7 CD PRO A 1 -3.214 -15.444 1.778 1.00 0.00 C
ATOM 8 HN1 PRO A 1 -3.192 -15.801 1.284 1.00 0.00 H
ATOM 9 HN2 PRO A 1 -3.078 -16.696 0.086 1.00 0.00 H
ATOM 10 HA PRO A 1 -4.659 -15.440 -1.190 1.00 0.00 H
ATOM 11 HB1 PRO A 1 -5.414 -13.348 0.409 1.00 0.00 H
ATOM 12 HB2 PRO A 1 -5.873 -15.000 0.847 1.00 0.00 H
ATOM 13 HG1 PRO A 1 -3.573 -13.318 1.902 1.00 0.00 H
ATOM 14 HG2 PRO A 1 -4.733 -14.301 2.804 1.00 0.00 H
ATOM 15 HD1 PRO A 1 -2.215 -15.228 2.186 1.00 0.00 H
ATOM 16 HD2 PRO A 1 -3.599 -16.321 2.320 1.00 0.00 H
After relaxation:
ATOM 1 N 0AZ A 1 -3.139 -15.719 0.289 1.00 0.00
ATOM 2 CA 0AZ A 1 -4.096 -14.883 -0.427 1.00 0.00
ATOM 3 C 0AZ A 1 -3.392 -13.770 -1.191 1.00 0.00
ATOM 4 O 0AZ A 1 -2.247 -13.429 -0.896 1.00 0.00
ATOM 5 CB 0AZ A 1 -5.012 -14.337 0.672 1.00 0.00
ATOM 6 CG 0AZ A 1 -4.140 -14.259 1.879 1.00 0.00
ATOM 7 CD 0AZ A 1 -3.211 -15.439 1.776 1.00 0.00
ATOM 8 OD1 0AZ A 1 -3.499 -12.987 1.901 1.00 0.00
ATOM 9 1H 0AZ A 1 -3.193 -15.801 1.284 1.00 0.00
ATOM 10 2H 0AZ A 1 -3.079 -16.696 0.086 1.00 0.00
ATOM 11 HA 0AZ A 1 -4.659 -15.440 -1.190 1.00 0.00
ATOM 12 1HB 0AZ A 1 -5.415 -13.348 0.407 1.00 0.00
ATOM 13 2HB 0AZ A 1 -5.873 -14.999 0.846 1.00 0.00
ATOM 14 1HG 0AZ A 1 -4.734 -14.302 2.803 1.00 0.00
ATOM 15 1HD 0AZ A 1 -2.212 -15.219 2.180 1.00 0.00
ATOM 16 2HD 0AZ A 1 -3.591 -16.316 2.321 1.00 0.00
ATOM 17 HD1 0AZ A 1 -2.583 -13.113 1.609 1.00 0.00
It’s quite weird, anyway…

[Anonymous]

It’s your hydrogen naming convention. (Or rather, it’s Rosetta naming conventions.) As mentioned, Rosetta uses a heuristic based on the presence/absence of atoms to determine how to interpret the residues.

You have HD1 and HD2 atoms on your delta carbon. Rosetta uses 1HD and 2HD as the names of those atoms, causing it to discard those atoms — except, the hydroxylated proline variant uses HD1 to represent the hydrogen on the oxygen attached to the delta carbon. Rosetta sees that you have that extra atom and assumes that’s because you’re inputting a hydroxyproline that’s just missing the oxygen. It then rebuilds the residue to fill in the missing oxygen, which causes the HD1 atom to be moved to be attached to the hydroxyl.

Remove (just delete) the HD1 proline lines from your input file and you should be fine. Rosetta will rebuild the missing atoms in the correct position.

[Author]

Posts